Alzheimer's Diagnosis: Leukocyte Multigene Signatures.

NINCDS-ADRDA和DSM-IV标准目前被广泛用于诊断可能的阿尔茨海默病（AD）。这些临床标准有许多局限性，包括诊断缺乏特异性和敏感性，即使在学术研究中心也有大约10%的错误率。此外，基于认知功能的诊断只能在术后进行，此时可能抑制AD发展或延迟其进展的药物可能无效。目前正在开发的成像和生物标志物诊断方法在它们分别需要高度专业化的设备以及特异性和灵敏度方面具有额外的缺点，因此可能对早期筛查没有用。该提议被设计为基于转录的白色血细胞（白细胞）RNA的高密度微阵列测量来产生用于开发AD患者的生物分类的初步数据。这一提议背后的基本原理基于两个数据来源：1）当前的科学文献，其中有越来越多的证据表明AD患者表现出免疫和其他反应，这些反应可以在循环外周白细胞中改变的基因表达水平上检测到。白细胞中一些单个基因的mRNA转录本的定量已经证明了基因表达水平与AD的存在之间的关联。2)来自PI的微阵列研究的初步结果，调查精神分裂症男性的基因表达变化。这项表达研究的初步结果是惊人的：对外周血白细胞基因表达数据进行监督聚类分析，使用来自8名精神分裂症患者和5名匹配对照受试者的数千个基因的转录水平测量，结果将所有受试者分类为正确的组。这些结果提供的证据表明，替代组织可以成功地用于神经精神疾病的分类。这些观察结果构成了本研究假设和实验设计的基础。利用类似的微阵列策略，我们建议调查我们的中心假设：患有AD的个体在其外周血白细胞中表现出保守的基因表达水平模式，这与对照受试者外周血白细胞中的表达模式不同。该提案的具体目标是：1A。在本项目的两年期间，收集30名特征良好的AD患者和30名特征良好的年龄匹配的健康对照受试者的血液白细胞，1b。采用Affyssin微阵列技术同时测量AD患者和对照受试者血液中白细胞中转录的多达47，000个基因的表达水平。2.为了分析在本提案的特定目标之一下收集的白细胞基因表达数据集，使用分层聚类和监督学习算法来识别和验证区分AD受试者与正常健康对照的基因表达模式（多基因签名）。这项探索性研究的成功完成对于为更大规模的NIH R 01研究提供初步数据、增加受试者数量以及纳入纵向研究以评估AD临床前发展和进展期间的多基因表达变化非常重要。我们相信，我们建议开发的诊断技术可能最终形成临床检测的基础，该检测将是微创的，并且将具有识别AD患者的能力，并且在未来还可能提供重要的症状前和早期诊断信息。

The NINCDS-ADRDA and DSM-IV criteria are currently widely used for diagnosis of probable Alzheimer's disease (AD). These clinical criteria have a number of limitations, including lack of specificity and sensitivity in the diagnosis, and have an error rate of approximately 10% even in academic research centers. Furthermore, diagnosis based on cognitive function can only be made post symptomatically, at which time medications that may inhibit AD development or delay its progression will likely be ineffective. The imaging and biological marker diagnostic methods currently under development have additional drawbacks in terms of their need for highly specialized equipment, and specificity and sensitivity respectively, and thus may not be useful for early screening This proposal is designed to produce preliminary data for development of a biological classification of AD patients, based on high-density microarray measurement of transcribed white blood cell (leukocyte) RNA. The rationale behind this proposal is based on two sources of data: 1) Current scientific literature, in which there is growing evidence that individuals with AD exhibit immune and other responses, that can be detected at the level of altered gene expression in circulating peripheral leukocytes. Quantitation of the mRNA transcripts in leukocytes of a number of individual genes has demonstrated associations between gene expression levels and the presence of AD. 2) Preliminary results from a microarray study by the PI, investigating gene expression changes in men with schizophrenia. Initial results from this expression study have been striking: Supervised cluster analysis of peripheral leukocyte gene expression data, using transcript level measurements of thousands of genes from eight schizophrenic patients and five matched control subjects, resulted in a classification of all the subjects into their correct group. These results provide evidence to suggest that a surrogate tissue can be successfully employed for classification of a neuropsychiatric disease. These observations form the basis of the hypothesis and experimental design of this proposed study. Utilizing a similar microarray strategy, we propose to investigate our central hypothesis: that individuals suffering from AD exhibit a conserved pattern of gene expression levels in their peripheral blood leukocytes, which is distinct from the pattern of expression in peripheral blood leukocytes from control subjects. The specific aims of this proposal are; 1a. To collect blood leukocytes from thirty well characterized, AD patients and thirty well-characterized, healthy age-matched control subjects over the two-year period of this project, 1b. To employ Affymetrix microarray technology to measure simultaneously the expression levels of up to 47,000 genes transcribed in leukocytes derived from the blood of the AD patients and control subjects. 2. To analyze the leukocyte gene expression datasets collected under Specific Aim One of this proposal, using hierarchical clustering and supervised learning algorithms to identify and validate patterns of gene expression (multigene signatures) that differentiate AD subjects from normal healthy controls. The successful completion of this exploratory study will be important for providing preliminary data for larger scale NIH R01 studies, to increase subject numbers and to incorporate a longitudinal study to assess multigene expression changes during the preclinical development and progression of AD. We believe that the diagnostic technique we propose to develop may ultimately form the basis of a clinical assay that will be minimally invasive, and will have the capacity to identify AD sufferers, and in the future may also provide important pre-symptomatic and early stage diagnostic information.